Conventional multimedia content such as video may be enhanced with haptic feedback. However, such enhancements typically require embedding information that encodes the haptic feedback directly into the video. Thus, these techniques are unable to provide real-time haptic feedback on pre-existing video that has not already been annotated with the information that encodes the haptic feedback.
Furthermore, conventional media playback systems do not leverage sensor information that indicates a motion of an object that appears in a scene in the video to generate haptic feedback. Additionally, estimating an acceleration of an object based on video information may be inaccurate when the object is travelling at high speeds due to blur and other artifacts, while estimating an acceleration of the object using information from a sensor attached to the object may be inaccurate at low speeds.